DESCRIPTION: Clinical literature often mentions telephone use as one of the most stressful communication situations for the 1 percent of the population who stutter. We propose a device to modify stuttered speech spoken into a telephone. It uses phonetic classification technology to detect certain types of dysfluencies. The dysfluencies will be removed and real-time time-scale modification will be used to seamlessly lengthen fluent portions of the utterances to compensate. Listeners will assess the quality and intelligibility of dysfluent (unprocessed) and fluency-enhanced (processed) speech in listening experiments. The proposed device is a tool that persons who stutter may choose to use when they need to talk on the telephone with greater fluency. In addition to modifying speech, the device might also modestly reduce speaker anxiety, a commonly accepted factor in fluency improvement. It will prove especially useful for those whose employment depends on telephone use. Its value could also extend to international markets, and the device could be modified for other electronically mediated communication, such as public address systems. PROPOSED COMMERCIAL APPLICATION: Surveys have shown that employers' attitudes about the way employees sound on the telephone affect employability and advancement; the effect is negative for persons who stutter. If successful, the proposed device will enable such persons to function more effectively in their work. The device may also qualify as a "reasonable accommodation" by the employer under the Americans with Disabilities Act.